Project 2. Project Summary. This Project 2 is a renewal of Project 3 in our current Conte Center. It has Ralph Adolphs (overall Center Director) and Cendri Hutcherson (formerly post-doc on current Project 3, now faculty) as co-PIs and includes a modest subcontract to the University of Toronto, Canada (Hutcherson). Its overall goal is to understand how social inference and context (the focus theme of this Conte Center) guide social decision-making, with a specific focus on prosocial (altruistic) decisions. A counterpart to this is Project 3, which will investigate how social inference and context guide decisions related to social threat. Project 2 will focus on how we represent the internal states of another person (social inference representations, Aim 1), how such representations are then used to make altruistic social decisions (Aim 2), and where in these processes we find individual differences that may correlate with measures like autistic traits, empathy, or social network size (Aim 3). As with Projects 1 and 3 in this Conte Center renewal, this Project 2 focuses on fMRI studies in healthy individuals. Also as with Projects 1 and 3, this Project 2 has links to Aims under other Projects that offer a complementary approach. Specifically, it features computational models (drift-diffusion models) that will also be leveraged in Project 3, and it will develop a battery of social inference tasks (versions of the RDoC-listed ?why/how? task) that will also be administered to patients with focal lesions of the prefrontal cortex under Project 5. As well, we plan to make specific comparisons between the neural systems for social inference revealed under this Project 2 with social neuroscience tasks such as the ?why/how? task, and the systems revealed with computational fMRI investigated for observational learning in Project 1 (using cross-task decoding in overlapping subjects). The strong links between Project 2 and others is reflected in its personnel: it lists PIs from other Projects (O'Doherty, Tranel, Camerer) and shares post-docs with other Projects. The first two Aims will develop tasks for two focused fMRI studies that will each be conducted in 100 healthy participants recruited through Cores 2 and 3. A subset of 30 of these will be retested in future years to investigate stability over time. Aim 1 will develop six tasks that probe how specific contexts modulate social inference: through task set, for facial expressions or actions, for social or nonsocial events, as a function of cognitive load, and whether the person we are observing seems to be in need or to have merit. This latter need/merit context factor is further developed in relation to altruistic decision-making in Aim 2, which expands a task and approach we recently published, using drift-diffusion modeling to better understand how specific parameters influence the temporal evolution of the choice process.